168
B a k e r , On some diophantine inequalities involving primes
The highest
common factor of integers ra, n, . .
k
is denoted by (m, ra, . . .,
k ) .
For each pair of integers
m , q > 0
we define
and put
Cq(m) = J S
3)
= J ( 17
C?(m) (?(? ))
i = i u - i
The series converges, as will be clear later from the proof of Lemma 7.
By c19 c2, . . . we shall mean positive absolute constants unless the contrary is
specifically stated, [x] denotes the integral p a rt of x and || x || denotes th e distance of x
from the nearest integer taken positively.
3. Lemmas
We now give fourteen lemmas prelim inary to the proof of th e theorem.
Lemma 1. For each integer
b
4
=
0 and each
a
on
W a Q
we have
(6)
I
S(bcc) - I(bfi)CQ(b)(cpM) - 1 1 < c, I
!
where fi — a — a/q and cly c2 > 0 depend only on H .
Proof. The proof follows as in [4] (see p. 188) on noting th a t the conclusion of
Lem m a 4 of [4] can be sharpened to
:
*
\ f
h i '
}
i
¿ t ip ‘ h iIt
where c3y cx > 0 depend only on oc (see [3], Satz 8. 3, p. 144). The presence of |
b
| on
the right of (6) results from the inequalities
N \ 1 — e(bfi) | ^ 2 j t N |
bf i
2
tz
Dostları ilə paylaş: